TSG101 depletion dysregulates mitochondria and PML NBs, triggering MAD2-overexpressing interphase cell death (MOID) through AIFM1-PML-DAXX pathway

Abstract

Overexpression of mitotic arrest deficiency 2 (MAD2/MAD2L1), a pivotal component of the spindle assembly checkpoint (SAC), resulted in many types of cancer. Here we show that the depletion of tumor susceptibility gene 101 (TSG101), causes synthetic dosage lethality (SDL) in MAD2-overexpressing cells, and we term this cell death MAD2-overexpressing interphase cell death (MOID). The induction of MOID depends on PML and DAXX mediating mitochondrial AIFM1-release. MAD2, TSG101, and AIF-PML-DAXX axis regulate mitochondria, PML nuclear bodies (NBs), and autophagy with close inter-dependent protein stability in survival cells. Loss of C-terminal phosphorylation(s) of TSG101 and closed (C-)MAD2-overexpression contribute to induce MOID. In survival cells, both MAD2 and TSG101 localize at PML NBs in interphase, and TSG101 Y390 phosphorylation is required for localization of TSG101 to PML NBs. PML release from PML NBs through PML deSUMOylation contributes to induce MOID. The post-transcriptional/translational cell death machinery and the non-canonical transcriptional regulation are intricately linked to MOID, and ER-MAM, may serve as a crucial intersection for MOID signaling.

Fig. 1. Depletion of TSG101 showed synthetic dosage lethality (SDL) in MAD2-overexpressing cells.

A TUNEL assay. HeLa cells were transfected with the indicated siRNAs (Table S1) and/or constructs (Table S2). Cells were cultured for 96 h at 37°C. DNA fragmentation was detected by the TUNEL assay, and samples underwent indirect fluorescence microscopy using anti-Flag as a primary antibody to sort out Flag-MAD2 overexpressing cells (see MATERIALS AND METHODS). Scale bar, 10 μm. B Western blot analysis of total lysates of HeLa cells transfected with the indicated siRNAs and/or constructs. Cells were cultured for 96 h, collected, lysed, and immunoblotted with the indicated antibodies (Table S3). C A histogram summarizing TUNEL assay results of (A) (see MATERIALS AND METHODS). The mean percentages (±SD) of TUNEL-positive cells were shown. D A histogram summarizing TUNEL assay results (Images are shown in Fig. S2E and S2F). The mean percentages (±SD) of TUNEL-positive cells were shown. E TUNEL assay. Scale bar, 10 μm. (F) TUNEL assay. Scale bar, 10 μm. G A histogram summarizing TUNEL assay results of (E) and (F). The mean percentages (± SD) of TUNEL-positive cells were shown. (****) P < 0.0001 compared with 1st column from left in each cell line (Student’s t-test). H (Left) Colony outgrowth assay. MOID reduces the viability of HSkMCs. (Right) Colony counts and area (%) normalized to the control (1^st column from left) are shown summarizing the results of (Left). The mean percentages (± SEM) were shown. I Representative images of zebrafish xenografts of HSkMCs (Table S5). Scale bar, 500 μm. J HSkMCs-grown area and volume (%) are shown summarizing the results of (I) (Table S5). (Left) A line chart from 2-6 dpi. Data of each sample at 2 dpi was normalized to 100%. The mean percentages (± SEM) were shown. (****) P < 0.0001, (***) P < 0.001, and (**) P < 0.01 (one-way ANOVA). (Right) A histogram of 6 dpi. Data of each sample at 2 dpi was normalized to 100%. The mean percentages (± SEM) were shown. (K-N) Mouse xenografts implanting HeLa cells (see MATERIALS AND METHODS; Table S5). (K) Representative images with scale bar, 10 mm. L Tumor volume (mm²) at post engraftment (days). ⊠, invisible size of tumor. The mean percentages (± SEM) were shown. (****) P < 0.0001 and (**) P < 0.01 (one-way ANOVA). M Tumor weight. The mean percentages (± SEM) were shown. N Body weight. The mean percentages (± SEM) were shown.

Fig. 2. MOID induction depends on PML-DAXX function concomitant with mitochondrial AIFM1-release, but the inductive or suppressive role of PML-DAXX during MOID is TSG101-dependent.

A and B Peptide set enrichment analysis (PSEA; see MATERIALS AND METHODS) of Flag-MAD2 interacting peptides was performed for 293 T (C) and HeLa+293 T (D) samples. P- and FDRq-values and NES were indicated. C Peptides interacting with MAD2 were identified from Mass-spec in both 293 T and HeLa, which were enriched in 112 and 117 pathways after PANTHER19.0 pathway analysis, respectively. 10 genes were identified from both 293 T and HeLa, which were annotated as Apoptosis signaling pathway (P00006) (MATERIALS AND METHODS). D A table showing that AIFM1 is found among 10 genes identified as C. E and F An anti-AIFM1 (green) rabbit polyclonal antibody (Table S3) was used to detect the release of AIFM1 from mitochondria marked with MitoTracker (red) during MOID. Arrows, damaged nuclei. Arrowheads, low DAPI nuclei. Scale bar, 10 μm. G Quantitated immunofluorescence signals of AIFM1 in nuclear shown in E and F (see MATERIALS AND METHODS). Signals were normalized to those of each Vector control (Ctrl), and the mean percentages (±SEM) are shown. H A histogram summarizing the pattern of MitoTracker signals shown in E and F (see MATERIALS AND METHODS). The mean percentages (± SD) of each pattern were shown. I Subcellular fractions isolated from 293 T and HeLa cells after the transfection with the indicated siRNAs and/or constructs. N, nuclear; Mc, crude mitochondria. STS, staurosporin (1.0 μM) control (see MATERIALS AND METHODS). Lamin A/C was used as a nuclear marker, TOM20 as a mitochondria marker. Note that we could not detect AIFM1 bands even with staurosporine (1.0–100 μM) control in cytosol fraction, presumably due to diluted conditions during sample preparation (data not shown). J A histogram summarizing TUNEL assay results as described in Fig. 1C (see MATERIALS AND METHODS). The mean percentages (± SD) of TUNEL-positive cells were shown. iVec indicates CRISPRi vector control. koVec indicates CRISPR KO vector control. K A histogram summarizing TUNEL assay results as described in Fig. 1C (see MATERIALS AND METHODS). The mean percentages (± SD) of TUNEL-positive cells were shown. iVec indicates CRISPRi vector control. koVec indicates CRISPR KO vector control.

Fig. 3. Loss of C-terminal phosphorylations of TSG101 and C-MAD2-overexpression contribute to induce MOID.

A Schematic figures of each TSG101 construct used in this study. B Western blot analysis of total lysates of HeLa cells transfected with the indicated siRNAs and/or constructs. Cells were cultured for 48 h, collected, lysed, and immunoblotted with the indicated antibodies (Table S3). Asterisk indicates background signals. C Loss of C-terminal phosphorylations of TSG101 contributes to induce MOID. 293 T cells were transfected with the indicated siRNAs and/or constructs. Cells were cultured for 96 h at 37°C. DNA fragmentation was detected by the TUNEL assay (see MATERIALS AND METHODS). Scale bar, 10 μm. D A histogram summarizing TUNEL assay results shown in C. The mean percentages (± SD) of TUNEL-positive cells were shown. E C-MAD2-overexpression is required to induce MOID. HeLa cells were transfected with the indicated siRNAs and/or constructs. Cells were cultured for 96 h at 37°C. DNA fragmentation was detected by the TUNEL assay as described in Fig. 1A (see MATERIALS AND METHODS). Scale bar, 10 μm. F A histogram summarizing TUNEL assay results shown in E. The mean percentages (± SD) of TUNEL-positive cells were shown. G O-MAD2-overexpression does not induce MOID and maintains PML SUMOylation. 293 T cells were transfected with the indicated constructs, and pellets were collected 96 h after the transfection, which is the timing when TUNEL-positive MOID cells were observed. Proteins in 5% of the total cell lysates (Input) and immunoprecipitates (IP) were detected by western blot analysis using the indicated antibodies. H MOID induced by C-MAD2-overexpression promotes multimerization and SUMOylation of C-MAD2 at the later stage of MOID, while overexpressed O-MAD2 that blocks MOID is degraded at the same timing. 293 T cells were transfected with the indicated constructs, and pellets were collected 120 h after the transfection (24 h after the timing of (G) sample). Proteins in 5% of the total cell lysates (Input) and immunoprecipitates (IP) were detected by western blot analysis using the indicated antibodies. I Majority of overexpressed O-MAD2 binds Y390-intact TSG101 WT, while overexpressed C-MAD2 binds TSG101 Y390F mutant. Immunoprecipitation assay (see MATERIALS AND METHODS). 293 T cells were transfected with the indicated constructs, and pellets were collected 96 h after the transfection, which is the timing when TUNEL-positive MOID cells were observed. Proteins in 1.6% of the total cell lysates (Input) and immunoprecipitates (IP) were detected by western blot analysis using the indicated antibodies. Asterisk (*) indicates putative posttranslationally modified TSG101 band. Hash (#) indicates non-specific bands. J A histogram summarizing TUNEL assay results corresponding to the samples shown in I using untagged-TSG101. The mean percentages (± SD) of TUNEL-positive cells were shown. K Immunoprecipitation assay performed as I (see MATERIALS AND METHODS), but using HA-TSG101. Proteins in 3.0% of the total cell lysates (Input) and immunoprecipitates (IP) were detected by western blot analysis using the indicated antibodies. L A histogram summarizing TUNEL assay results corresponding to the samples shown in K using HA-TSG101. The mean percentages (± SD) of TUNEL-positive cells were shown.

Fig. 4. MAD2 colocalizes with PML at PML NBs, and PML release from PML NBs through PML deSUMOylation contributes to induce MOID.

A PML constructs used in this study. All the PML constructs were fused to an N-terminal HA tag and RNAi-resistant. B Western blot analysis using 293 T total cell lysates. Cells were cultured and harvested 96 h after transfection with the indicated constructs. C Representative images of the in vivo SUMOylation assay with the HA-PML constructs (WT and mutants) using 293 T cells (see MATERIALS AND METHODS). D DeSUMOylated PML delta coiled-coil (dCC) mutant did not suppress MOID induction with Flag-MAD2 overexpression and single PML depletion, and constitutively induces MOID. TUNEL assay of 293 T cells was performed as described in Fig. 1A (see MATERIALS AND METHODS).Scale bar, 10 μm. E A histogram summarizing TUNEL assay results shown in D. The mean percentages (± SD) of TUNEL-positive cells were shown. F In vivo SUMOylation assay of HA-PML using 293 T cells (see Figure S8A; MATERIALS AND METHODS). G In vivo SUMOylation assay of endogenous PML using 293 T cells (see MATERIALS AND METHODS). H Immunofluorescence images of cells transfected with the indicated siRNA and/or constructs are shown (see MATERIALS AND METHODS). Arrows, damaged nuclei. Arrowheads, low DAPI nuclei. I The formation of PML NB is induced by only MAD2 overexpression, and significantly reduced by TSG101 depletion, but is further abolished under MOID conditions. A histogram summarizing the quantitation of PML signals in interphase of H. Sample numbers correspond to the ones shown in Fig. 4H and Tabe S5. Signals were normalized with Ctrl cells (1st column from left), and the mean percentages (±SEM) are shown. J A histogram summarizing the quantitation of nuclear MAD2 signals including interphase foci of H. Sample numbers correspond to the ones shown in Fig. 4H and Tabe S5. Signals were normalized with Ctrl cells (1st column from left), and the mean percentages (±SEM) are shown. K A histogram summarizing the Pearson correlation coefficient of the colocalization between MAD2 nuclear foci and PML signals of H. Sample numbers correspond to the ones shown in Fig. 4H and Tabe S5. The mean values (±SEM) are shown.

Fig. 5. Transcriptome RNA-seq analysis of MOID.

A Scheme of cell treatments and HeLa samples’ collection for RNA-seq. Venn diagram shows the gene number of shared or unique expressed genes in each group (submitted sample number = 2). B Histogram of the gene number of both up or downregulated genes in the matched MOID group compared with the control group. C Volcano plot of differentially expressed genes in the MOID group compared with the control group. Cut-off values, p ≤ 0.05, FC ≥ 1.5. D–I Gene set enrichment analysis (GSEA) of genes’ set that hit the following keywords: D autophagy, E oxidation, F apoptotic, G cell-adhesion, H cilium, I DNA damage and/or DNA repair. P- and FDRq-values and NES were indicated. J Gene Ontology (GO) enrichment analysis showing top 25 biological processes of 76 genes affected in the MOID group. Orange rectangle indicates apoptotic process. K Evaluation of differentially expressed genes (with most significant P value/each process and reads counts ≥ 10) of top 25 biological processes in GO enrichment analysis by RT-qPCR (n ≥ 3 replicates). The mean ± SEM. is shown. L Evaluation of differentially expressed genes (with most significant P value and reads counts ≥ 10) in apoptotic related gene sets as shown in (J, orange rectangle) by RT-qPCR (n ≥ 3 replicates). The mean ± SEM. is shown. M Reactome analysis showing 25 pathways of 11 genes affected in the MOID group. Violet rectangle indicates HDACs/DNMT3s-related pathway. Green rectangle indicates ATM/ATR/BRCA1-related pathway. N Evaluation of differentially expressed genes (with most significant P value and reads counts ≥ 10) in ATM/ATR/BRCA1 related gene sets as shown in (M, green rectangle) by RT-qPCR (n ≥ 3 replicates). The mean ± SEM. is shown.

Fig. 6. ROS and autophagy are activated in MOID, and mitochondrial associated membrane (MAM) and PML localized at ER-MAM can be a crucial intersection for MOID signaling.

A ROS detection through fluorescent DCF using DCFH-DA probe (see MATERIALS AND METHODS). B A histogram summarizing ROS detection (fluorescence DCF intensity) as shown in A. HeLa cells were transfected with the indicated constructs, and pellets were collected 96 h after the transfection as TUNEL-positive MOID cells were observed. The mean percentages (± SD) were shown (see MATERIALS AND METHODS, Statistical Analysis). C Quantitation of interphase intracellular signals of anti-BECLIN1 (BECN1) in interphase. HeLa cells were transfected as samples [1]-[6] in A and B and cultured at 37 °C for 96 hours, which is the timing when TUNEL-positive MOID cells were observed. Samples were observed with an indirect fluorescence microscope using an anti-BECN1 antibody (Tabe S3). D Quantitation of interphase intracellular signals of anti-LC3B in interphase. HeLa cells were transfected as samples [1]-[6] in (A and B) and cultured at 37 °C for 96 hours, which is the timing when TUNEL-positive MOID cells were observed. Samples were observed with an indirect fluorescence microscope using an anti-LC3B antibody (Tabe S3). E Quantitation of interphase intracellular signals of MitoTracker in interphase (see MATERIALS AND METHODS). HeLa cells were transfected as samples [1]-[6] in (A and B) and cultured at 37 °C for 96 hours, which is the timing when TUNEL-positive MOID cells were observed. F Pearson correlation coefficient between fluorescence signals of MitoTracker and anti-BECN1 in interphase (see MATERIALS AND METHODS). G Pearson correlation coefficient between fluorescence signals of MitoTracker and anti-LC3B in interphase (see MATERIALS AND METHODS). H-M Immunofluorescence signals of ER-MAM markers are associated with the MOID. Representative immunofluorescence images and histograms summarizing quantitation of interphase intracellular signals of (H and I) anti-IP3R3, (J and K) anti-SIGMA1R, and (L and M) anti-phospho-IP3R1 (Ser 1756) (see MATERIALS AND METHODS). N Western blot analysis of HA-PML WT and mutants (erPML and gbPML). O and P erPML mutant failed to rescue this MOID induction with Flag-MAD2 overexpression and single PML depletion. A histogram summarizing TUNEL assay of (O) HeLa and (P) 293 T cells performed as described in Fig. 1C (see MATERIALS AND METHODS). The mean percentages (± SD) of TUNEL-positive cells were shown. Q A hypothetical model depicting that PML-MAD2 interaction at PML NBs and formation of PML NBs regulates MOID. (Left) In survival interphase, PML, DAXX, MAD2, and TSG101 maintain a close dependence and contribute mutually to protein stability sustaining inter-organelle integrity. Note that previous reports suggested that PML can be auto-SUMOylated [40, 41, 68] and/or SUMOylated by PIAS1 SUMO E3 ligase [69]. O-MAD2-specific binding of TSG101 and TSG101 function inside PMN NB are hypothetical. Black arrows indicate activatory or inhibitory direction based on our results of immunofluorescence and protein stability assay or other reports. Black double-head arrows indicate colocalization based on our immunofluorescence results. Blue arrows indicate inhibitory or interactive relationships based on our immunoprecipitation and/or in vivo SUMOylation assays. (Right) The series of our results propose at least 2 steps of MOID induction process (1): “Priming” and (2) “Releasing”. Mechanisms and roles of MAD2 SUMOylation/multimerization and autophagy are open questions. Transcriptional roles of AIFM1 and PML-DAXX collaboration are unknown. See text for other details.